Datasets:

ArXiv:
Elron commited on
Commit
4ebbd89
1 Parent(s): 9e7eeeb

Upload metrics.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. metrics.py +41 -2
metrics.py CHANGED
@@ -8,6 +8,7 @@ import evaluate
8
  import nltk
9
  import numpy
10
 
 
11
  from .operator import (
12
  MultiStreamOperator,
13
  SingleStreamOperator,
@@ -60,7 +61,13 @@ class GlobalMetric(SingleStreamOperator, Metric):
60
 
61
  refs, pred = instance["references"], instance["prediction"]
62
 
63
- instance_score = self._compute([refs], [pred])
 
 
 
 
 
 
64
  instance["score"]["instance"].update(instance_score)
65
 
66
  references.append(refs)
@@ -355,8 +362,27 @@ class Bleu(HuggingfaceMetric):
355
  scale = 1.0
356
 
357
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
358
  class CustomF1(GlobalMetric):
359
  main_score = "f1_micro"
 
360
 
361
  @abstractmethod
362
  def get_element_group(self, element):
@@ -391,6 +417,10 @@ class CustomF1(GlobalMetric):
391
  assert len(references) == len(predictions), (
392
  f"references size ({len(references)})" f" doesn't mach predictions sise ({len(references)})."
393
  )
 
 
 
 
394
  groups_statistics = dict()
395
  for references_batch, predictions_batch in zip(references, predictions):
396
  grouped_references = self.group_elements(references_batch)
@@ -418,6 +448,7 @@ class CustomF1(GlobalMetric):
418
  groups_statistics[group]["recall_denominator"] += rd
419
 
420
  result = {}
 
421
  pn_total = pd_total = rn_total = rd_total = 0
422
  for group in groups_statistics.keys():
423
  pn, pd, rn, rd = (
@@ -426,13 +457,21 @@ class CustomF1(GlobalMetric):
426
  groups_statistics[group]["recall_numerator"],
427
  groups_statistics[group]["recall_denominator"],
428
  )
429
- result[f"f1_{group}"] = self.f1(pn, pd, rn, rd)
430
  pn_total, pd_total, rn_total, rd_total = pn_total + pn, pd_total + pd, rn_total + rn, rd_total + rd
 
 
 
 
431
  try:
432
  result["f1_macro"] = sum(result.values()) / len(result.keys())
433
  except ZeroDivisionError:
434
  result["f1_macro"] = 1.0
435
 
 
 
 
 
 
436
  result[f"f1_micro"] = self.f1(pn_total, pd_total, rn_total, rd_total)
437
  return result
438
 
 
8
  import nltk
9
  import numpy
10
 
11
+ from .dataclass import InternalField
12
  from .operator import (
13
  MultiStreamOperator,
14
  SingleStreamOperator,
 
61
 
62
  refs, pred = instance["references"], instance["prediction"]
63
 
64
+ try:
65
+ instance_score = self._compute([refs], [pred])
66
+ except:
67
+ instance_score = {"score": None}
68
+ if isinstance(self.main_score, str) and self.main_score is not None:
69
+ instance_score[self.main_score] = None
70
+
71
  instance["score"]["instance"].update(instance_score)
72
 
73
  references.append(refs)
 
362
  scale = 1.0
363
 
364
 
365
+ class MatthewsCorrelation(HuggingfaceMetric):
366
+ metric_name = "matthews_correlation"
367
+ main_score = "matthews_correlation"
368
+ str_to_id: dict = InternalField(default_factory=dict)
369
+
370
+ def get_str_id(self, str):
371
+ if str not in self.str_to_id:
372
+ id = len(self.str_to_id)
373
+ self.str_to_id[str] = id
374
+ return self.str_to_id[str]
375
+
376
+ def compute(self, references: List[List[str]], predictions: List[str]) -> dict:
377
+ formatted_references = [self.get_str_id(reference[0]) for reference in references]
378
+ formatted_predictions = [self.get_str_id(prediction) for prediction in predictions]
379
+ result = self.metric.compute(predictions=formatted_predictions, references=formatted_references)
380
+ return result
381
+
382
+
383
  class CustomF1(GlobalMetric):
384
  main_score = "f1_micro"
385
+ classes = None
386
 
387
  @abstractmethod
388
  def get_element_group(self, element):
 
417
  assert len(references) == len(predictions), (
418
  f"references size ({len(references)})" f" doesn't mach predictions sise ({len(references)})."
419
  )
420
+ if self.classes is None:
421
+ classes = set([self.get_element_group(e) for sublist in references for e in sublist])
422
+ else:
423
+ classes = self.classes
424
  groups_statistics = dict()
425
  for references_batch, predictions_batch in zip(references, predictions):
426
  grouped_references = self.group_elements(references_batch)
 
448
  groups_statistics[group]["recall_denominator"] += rd
449
 
450
  result = {}
451
+ num_of_unknown_class_predictions = 0
452
  pn_total = pd_total = rn_total = rd_total = 0
453
  for group in groups_statistics.keys():
454
  pn, pd, rn, rd = (
 
457
  groups_statistics[group]["recall_numerator"],
458
  groups_statistics[group]["recall_denominator"],
459
  )
 
460
  pn_total, pd_total, rn_total, rd_total = pn_total + pn, pd_total + pd, rn_total + rn, rd_total + rd
461
+ if group in classes:
462
+ result[f"f1_{group}"] = self.f1(pn, pd, rn, rd)
463
+ else:
464
+ num_of_unknown_class_predictions += pd
465
  try:
466
  result["f1_macro"] = sum(result.values()) / len(result.keys())
467
  except ZeroDivisionError:
468
  result["f1_macro"] = 1.0
469
 
470
+ amount_of_predictions = pd_total
471
+ if amount_of_predictions == 0:
472
+ result["in_classes_support"] = 1.0
473
+ else:
474
+ result["in_classes_support"] = 1.0 - num_of_unknown_class_predictions / amount_of_predictions
475
  result[f"f1_micro"] = self.f1(pn_total, pd_total, rn_total, rd_total)
476
  return result
477